Interactions between vitamin D and IGF-I: from physiology to clinical practice

Authors

  • Pietro Ameri,

    1. Division of Endocrinology, Department of Medicine, New York University School of Medicine, New York, NY, USA
    2. Unit of Internal Medicine, Department of Internal Medicine and Medical Specialties, IRCCS-AOU San Martino-IST, University of Genova, Genova, Italy
    Search for more papers by this author
  • Andrea Giusti,

    1. Department of Geriatrics and Musculoskeletal Sciences, E.O. Galliera Hospital, Genova, Italy
    Search for more papers by this author
  • Mara Boschetti,

    1. Unit of Endocrinology & Centre of Excellence for Biomedical Research, Department of Internal Medicine and Medical Specialties, IRCCS-AOU San Martino-IST, University of Genova, Genova, Italy
    Search for more papers by this author
  • Giovanni Murialdo,

    1. Unit of Internal Medicine, Department of Internal Medicine and Medical Specialties, IRCCS-AOU San Martino-IST, University of Genova, Genova, Italy
    Search for more papers by this author
  • Francesco Minuto,

    1. Unit of Endocrinology & Centre of Excellence for Biomedical Research, Department of Internal Medicine and Medical Specialties, IRCCS-AOU San Martino-IST, University of Genova, Genova, Italy
    Search for more papers by this author
  • Diego Ferone

    Corresponding author
    1. Unit of Endocrinology & Centre of Excellence for Biomedical Research, Department of Internal Medicine and Medical Specialties, IRCCS-AOU San Martino-IST, University of Genova, Genova, Italy
    • Division of Endocrinology, Department of Medicine, New York University School of Medicine, New York, NY, USA
    Search for more papers by this author

Correspondence: Diego Ferone, Endocrinology Unit (DiMI), University of Genova, Viale Benedetto XV, 6, Genova 16132, Italy. Tel.: +39 010 353 7946; Fax: +39 010 353 8977; E-mail: ferone@unige.it

Summary

The interplay between vitamin D and IGF-I is complex and occurs at both endocrine and paracrine/autocrine levels. Vitamin D has been shown to increase circulating IGF-I and IGFBP-3, with the consistent finding of a positive correlation between vitamin D and IGF-I serum values in population-based cohorts of healthy subjects. The modulation of IGF-I and IGFBP-3 concentrations by vitamin D may impact recombinant human (rh) GH dosing for the treatment of GHD. It might also underlie some of the extra-skeletal beneficial effects ascribed to vitamin D. On the other hand, IGF-I stimulates renal production of 1,25-dihydroxyvitamin D, which increases calcium and phosphate availability in the body and suppresses PTH secretion. This effect is responsible for an altered calcium–phosphate balance in uncontrolled acromegaly and might also account for the improvement in bone metabolism associated with rhGH treatment in patients with GHD. Data on the paracrine/autocrine vitamin D−IGF-I interactions are abundant, but mostly not linked to one another. As a result, it is not possible to draw a comprehensive picture of the physiological and/or pathological interrelations between vitamin D, IGF-I and IGF-binding proteins (IGFBP) in different tissues. A potential role of vitamin D action is related to its association with carcinogenesis, a paradigm being breast cancer. Current evidence indicates that, in breast tumours, vitamin D modulates the IGF-I/IGFBP ratio to decrease proliferation and increase apoptosis.

Introduction

Once considered just a nutrient important for bone health, vitamin D has been rediscovered as a pleiotropic hormone during the last three decades.[1] By regulating the expression of about 3% of the genome, vitamin D affects the functions of most cell types.[2, 3] Not surprisingly, this widespread activity entails interplay with other hormones, among which is IGF-I. As both vitamin D and IGF-I are largely expressed throughout the body and have a broad spectrum of effects, their interrelations are extremely complex. This short review will first focus on how vitamin D and IGF-I influence their respective circulating concentrations, as this interaction may have clinical implications and may underlie some of the reported associations of vitamin D with extra-skeletal outcomes. Vitamin D−IGF-I paracrine/autocrine interactions will also be discussed, with emphasis on breast cancer.

Regulation and action of vitamin D

After being synthesized in the skin (cholecalciferol) or ingested with supplements or certain foods (both ergocalciferol and cholecalciferol), vitamin D is hydroxylated to 25-hydroxyvitamin D [25(OH)D] by cytochrome P-450-dependent enzymes.[2-4] Although there is evidence for some 25-hydroxylation of vitamin D in other organs, the liver accounts for most 25(OH)D production. Another cytochrome P-450-dependent hydroxylase, CYP27B1 or 1α-hydroxylase, catalyses the conversion of 25(OH)D into 1,25-dihydroxyvitamin D [1,25(OH)2D or calcitriol], the primary hormonally active form of vitamin D.[3, 4] By binding to and activating the nuclear vitamin D receptor (VDR), 1,25(OH)2D modulates the transcription of thousands of genes.[1, 3] The retinoid X receptor and several cofactors cooperate with the 1,25(OH)2D-VDR complex to regulate gene expression[5]. Some vitamin D effects occur in minutes to hours, before gene transcription is substantially affected, and are mediated by a VDR associated with the plasma membrane of certain cells.[5] Examples of these nongenomic responses are the rapid intestinal absorption of calcium,[5] the induction of exocytosis in Sertoli cells,[6] and the secretion of insulin by pancreatic β-cells.[5]

The 1α-hydroxylation reaction is the point at which the vitamin D system branches in two (Fig. 1). It has long been known that renal tubular cells produce 1,25(OH)2D from 25(OH)D supplied by the vasculature. Most 1,25(OH)2D synthesized in the kidney leaves the organ and, by acting in a classical endocrine way, participates in bone metabolism and stimulates intestinal absorption of calcium and phosphate and renal reabsorption of calcium. Furthermore, it inhibits PTH release from the parathyroid glands.[1-3] Renal 1α-hydroxylase and thus the endocrine actions of vitamin D are tightly regulated by phosphate, PTH and fibroblast growth factor 23 (FGF23).[3, 4] This latter is principally secreted by osteocytes and is a potent inhibitor of 1α-hydroxylation of 25(OH)D in the kidney.[7] Excess phosphate is a major stimulus for FGF-23 production. It is now believed that the bulk of circulating 25(OH)D is taken up by extra-renal 1α-hydroxylase-expressing tissues, which make their own 1,25(OH)2D for paracrine and autocrine activity.[2, 8] Extra-renal 1α-hydroxylase is principally regulated in a tissue-specific manner by several factors,[4] such as cytokines.[9]

Figure 1.

The vitamin D system. Vitamin D is made in the skin or ingested with food. To be activated, it is hydroxylated twice: first to 25-hydroxyvitamin D [25(OH)D] and then to 1,25-dihydroxyvitamin D [1,25(OH)2D]. While the liver accounts for most 25-hydroxylation, the 1α-hydroxylation reaction occurs in virtually any tissue of the body. In physiological conditions, however, only the 1,25(OH)2D produced by the kidney enters the circulation to act in an endocrine way, while the hormone synthesized in extra-renal sites functions in an autocrine/paracrine manner.

Levels of 1α-hydroxylase and the VDR are lower and those of the vitamin D-metabolizing enzyme CYP24A1 are higher in invasive breast carcinomas compared with benign lesions, possibly because of the selection of tumour cells capable of escaping growth inhibition exerted by vitamin D.[10] In general, peripheral production of 1,25(OH)2D accounts for many extra-skeletal effects of vitamin D that would otherwise be difficult to explain based on vitamin D endocrine actions alone.[2, 3] In normal conditions, locally produced 1,25(OH)2D is used rapidly and does not leak into the circulation.[2] Hence, serum 1,25(OH)2D levels are only a measure of renal-derived vitamin D. In contrast, the concentration of 25(OH)D is a good indicator of the amount of vitamin D globally available, for both endocrine and paracrine/autocrine functions.

Regulation and action of IGF-I

IGF-I is transported to tissues via blood or is secreted by local cells.[11] These two sources dictate the modality of action of the hormone, either endocrine or paracrine/autocrine. As for vitamin D, the latter prevails.

In mice, removal of the hepatic contribution results in a 75% reduction in serum IGF-I levels, suggesting that about 75% of circulating IGF-I is produced by the liver.[12] Most IGF-I is carried in the bloodstream within a ternary complex with IGF-binding protein (IGFBP)-3 and acid-labile subunit (ALS), which are also secreted by hepatic cells.[11, 13] This complex stabilizes IGF-I in the circulation, reducing its clearance and thus prolonging the supply to target cells.[14] Peripheral organs also release some IGF-I into the systemic circulation, but their input is minor.[10] While GH is the main drive for hepatic IGF-I synthesis, in other sites tissue-specific factors are equally or even more important.[11] In addition, all six known IGFBPs regulate the availability of IGF-I to membrane receptors in a dynamic and complex manner.[11]

In mammals, local IGF-I cannot be substituted for liver-derived (endocrine) IGF-I for some specific effects: in addition to the well-known feedback inhibition of GH secretion from the pituitary gland, examples are maintenance of vascular resistance[13] and cardiac morphology and function.[15] In contrast, there are activities of IGF-I which primarily depend on the locally produced hormone, such as induction of mammary gland development.[16]

Relationship between circulating vitamin D and IGF-I

A positive correlation between serum concentrations of IGF-I, 25(OH)D and 1,25(OH)2D has been repeatedly demonstrated in healthy subjects.[17-20] Current evidence indicates that this association is at least in part causal. Treatment with cholecalciferol significantly increased circulating IGF-I in vitamin D-deficient children[21, 22] and adults (P. Ameri, A. Giusti, M. Boschetti, M. Bovio, C. Teti, G. Leoncini, D. Ferone, G. Murialdo, F. Minuto, Submitted). Consistently, Vdr−/− mice exhibit reduced IGF-I blood levels.[23] Serum IGFBP-3 was also found to be significantly higher after vitamin D supplementation.[21] On the other hand, short-term administration of both GH and IGF-I raised 1,25(OH)2D concentrations in healthy volunteers.[24, 25] The experimental data to explain how vitamin D and IGF-I reciprocally increase their blood levels are presented in the following two paragraphs. While there is no conclusive information about the mechanism(s) by which vitamin D modifies IGF-I and IGFBP-3 concentrations, it is well established that IGF-I stimulates the synthesis of 1,25(OH)2D in the kidney (Fig. 2).

Figure 2.

Schematic representation of vitamin D−IGF-I interactions. Vitamin D stimulates hepatic production of IGF-I and IGFBP-3. It is still not known how this effect occurs; however, experimental evidence suggests that vitamin D acts directly on liver cells. On the other hand, IGF-I stimulates the synthesis and activity of renal 1α-hydroxylase, thereby promoting vitamin D endocrine actions. The vitamin D and IGF-I systems also interact in an autocrine/paracrine fashion at the single tissue level.

Effects of vitamin D on IGF-I

No mechanistic study has yet addressed the effect of vitamin D on circulating IGF-I and IGFBP-3. Vitamin D is likely to act in the liver, as this organ accounts for most IGF-I and IGFBP-3 in the bloodstream. Based on the results of related studies, it can be hypothesized that vitamin D promotes liver production of IGF-I and IGFBP-3 by directly inducing the transcription of the relevant genes and/or by enhancing GH stimulation. Nonparenchymal hepatic cells (stellate, Kupffer, sinusoidal endothelial), rather than hepatocytes, might be the target of vitamin D, as they strongly express the VDR, respond to vitamin D in vitro[26, 27], and contribute to the pool of liver-derived circulating IGF-I and IGFBP-3.[28]

In mice knockout for steroid receptor co-activator 3 (SRC-3), a co-activator of nuclear receptors including the VDR, liver expression and thereby serum concentrations of IGFBP-3 are decreased; as a consequence, circulating IGF-I is also reduced because of enhanced clearance from the blood.[29] Therefore, vitamin D might modulate the hepatic synthesis of IGFBP-3 via SRC-3.

Although both the VDR[30] and CYP27B1[3] are expressed in the pituitary gland, vitamin D was found not to significantly influence GH secretion.[31] Moreover, vitamin D-deficient patients with secondary hyperparathyroidism have a normal GH response to GHRH + arginine.[32]

Finally, it is possible that vitamin D increases IGF-I concentrations by augmenting intestinal calcium absorption, as a high-calcium rescue diet has been reported to normalize IGF-I levels in Vdr−/− mice[23] and intake of calcium was positively associated with circulating IGF-I in humans.[33]

Effects of IGF-I on vitamin D

CYP2R1 is currently considered the principal enzyme responsible for 25-hydroxylation of vitamin D.[4] However, other hydroxylases may catalyse the reaction, such as CYP27A1. GH and IGF-I were shown to increase the endogenous activity of this enzyme in human hepatoblastoma cells.[34] However, this effect is likely to be negligible in vivo, as 25(OH)D concentrations do not vary even when GH and IGF-I change substantially, for example after starting treatment in patients with GH deficiency (GHD)[35] or acromegaly.[36]

In contrast, robust experimental evidence dating from the 1990s onwards indicates that IGF-I induces the synthesis and activity of renal 1α-hydroxylase.[24, 25, 37-39] GH also does so, but probably through local and endocrine IGF-I. Both GH and IGF-I increased the production of 1,25(OH)2D by cultured kidney cells, but the effect of GH was prevented by incubation with an antibody against IGF-1. Notably, IGF-I also stimulated the synthesis of 1,25(OH)2D by the placenta,[40] consistent with the significant positive correlation between IGF-I and 1,25(OH)2D observed in pregnant women.[41] This finding is particularly noteworthy as it adds to the understanding of the modulation of 1α-hydroxylation of 25(OH)D, which is uniquely not under the control of classic regulators (calcium, phosphate, PTH) during pregnancy.[42]

Paracrine/autocrine interactions between vitamin D and IGF-I

Although paracrine/autocrine interactions between the vitamin D and IGF-I systems have been described for many different tissues and cells, data are often insufficient to be pooled together in a comprehensive way. In this respect, investigations into the interplay between the vitamin D and IGF-I systems in the bone are typical. Despite the major clinical manifestations of vitamin D deficiency in the skeleton, the direct actions of vitamin D on bone are yet to be dissected.[43] As a consequence, although several authors have described interactions between vitamin D metabolites, IGF-I and IGFBPs that may regulate osteoblast and chondrocyte differentiation, proliferation and function,[44-47] it is presently difficult to integrate these results into a global picture.

Similarly, data suggest that vitamin D and IGFBPs act in concert to modulate insulin sensitivity,[18, 48] but the precise mechanisms remain elusive.

In contrast, vitamin D−IGF-I−IGFBP-3 interactions in breast cancer can be rather comprehensively summarized. IGF-I is central to the development of hyperplasia, precancerous lesions and eventually tumours of the mammary gland.[49, 50] Importantly, in mammary carcinogenesis locally produced IGF-I is far more important than the circulating hormone.[51] IGFBP-3 may oppose IGF-I tumourigenic action both by binding IGF-I and preventing it from interacting with its receptor and through IGF-I−independent antitumour activities.[52] Vitamin D was shown to decrease proliferation and stimulate apoptosis of nonneoplastic mammary epithelial cells,[53] breast cancer cells[54-56] and mouse mammary tumours[57] by antagonizing the mitogenic and anti-apoptotic effects of IGF-I and by stimulating the expression of IGFBP-3.

Other elements of the IGF-I systems were also favourably modulated by vitamin D in cultured breast cancer cells. In the oestrogen receptor-positive MCF-7 cell line, vitamin D, as well as two vitamin D analogues, enhanced the expression of IGFBP-5,[58, 59] which inhibits mammary proliferation.[60] In the oestrogen receptor-negative cell line, MDA MB 231, 1,25(OH)2D down-regulated the genes encoding IGFBP-2 and matrix metalloproteinases[59] (proteases involved in the cleavage and thereby inactivation of IGFBP-3[52]).

Hence, breast cancer may be considered as a model of paracrine/autocrine interplay between the vitamin D and IGF-I systems. Several caveats must be recognized, the main one being that this interaction may be not relevant in the clinical setting. Furthermore, the interrelations between vitamin D, IGF-I and IGFBPs may characterize only one or some breast cancer subtypes. In fact, it has recently been suggested that the vitamin D pathway is primarily important for the biology of a specific subset of mammary epithelial cells and their transformed counterpart.[61] Finally, the fact that IGFBP-3 may actually promote breast cancer cell survival under certain circumstances warrants mention.[62]

Relevance of vitamin D−IGF-I interactions for the clinical endocrinologist

Recombinant human GH (rhGH) is the mainstay of therapy of adult GHD. Current guidelines recommend that rhGH dosing regimens are individualized to achieve clinical benefit and normalize IGF-I levels while minimizing side effects.[63] It is common clinical experience that the IGF-I response to rhGH is not consistent, differing from patient to patient. Knowing the factors that account for this variability may help titrating rhGH. For instance, women need more rhGH than men to attain similar IGF-I values,[64] and among women, those on oral oestrogen require even higher doses.[65] By retrospectively examining a series of adult GHD patients on replacement therapy, we found that cases with higher concentrations of 25(OH)D were more likely to have serum IGF-I levels at least equal to the sex- and age-specific median in the normal population. In addition, they were more likely to be treated with lower amounts of rhGH (P. Ameri, A. Giusti, M. Boschetti, et al, Submitted). Therefore, assessment of vitamin D status might be indicated to better determine the dose of rhGH for treatment of adult patients with GHD.

The ability to stimulate renal 1α-hydroxylase by rhGH-induced IGF-I seems to be retained in both adult[66, 67] and paediatric GHD patients.[68, 69] A few studies have reported the opposite, but the discrepancy may be due to significant differences in doses and duration of rhGH therapy.[35, 70] Increased circulating 1,25(OH)2D, reduced PTH concentrations and increased calcium and phosphate availability, all resulting from enhanced 1α-hydroxylase activity in the kidney, may contribute to the improvement in bone metabolism obtained by long-term administration of rhGH to patients with postmenopausal osteoporosis[71] and patients with GHD.[66]

The induction of renal 1α-hydroxylase by IGF-I also explains some biochemical abnormalities of uncontrolled acromegaly. IGF-I−induced 1,25(OH)2D promotes the absorption of calcium and phosphate in the intestine and the reabsorption of calcium in the kidney. Moreover, IGF-I directly enhances the reabsorption of phosphate in the proximal renal tubule.[38] Consequently, compared with controlled disease, active acromegaly is characterized by significantly higher serum 1,25(OH)2D and by an increase in plasma calcium and phosphate and urinary calcium excretion, sometimes up to frankly abnormal values.[36, 72-74] On the other hand, overt hypercalcaemia is not a typical feature of acromegaly, suggesting the existence of some compensatory mechanism. FGF-23, secreted in response to phosphate accumulation, might play a role by feedback inhibiting renal 1α-hydroxylase. Two independent groups found that FGF-23 concentrations significantly decreased following the removal of pituitary GH-secreting adenomas, in parallel with a reduction in phosphate concentrations.[73, 74] In a third study, FGF23 values were also lower after treatment than before, but not to a significant extent.[36]

Are vitamin D−IGF-I interactions behind some of the magic of vitamin D?

Vitamin D deficiency has been related to a variety of adverse health consequences and diseases. Remarkably, IGF-I and IGFBP-3 have also been identified as correlates or predictors of many vitamin D-associated outcomes. For example, in the general population, hypovitaminosis D has been associated with cardiovascular and metabolic risk factors, such as arterial hypertension[75] and insulin resistance,[76] for which IGF-I values in the middle–upper normal range have been reported to be protective (as reviewed in[77]). Consistently, the risk of ischaemic heart or cerebrovascular disease has been inversely related to both vitamin D[75] and IGF-I levels.[77]

In a community-based sample of more than 6800 45-year-old British subjects, the lowest prevalence of metabolic syndrome was observed in participants with the highest concentrations of both 25(OH)D and IGF-I.[19] In this study, the positive correlation between 25(OH)D and IGF-I values was only present among participants with 75–85 nm 25(OH)D or less. As pointed out by the authors, this threshold interestingly corresponds to the cut-off at which vitamin D health benefits are maximized.[2] Diorio and colleagues reported that, in premenopausal women, vitamin D intake was negatively associated with mammographic breast density, a marker of breast cancer risk, and that the association was stronger in those with high serum IGF-I and IGFBP-3.[78] In elderly patients who had sustained hip fractures, a positive correlation was disclosed between 25(OH)D and IGFBP-3, both being significantly lower than in community controls.[79]

Therefore, it can be speculated that increases in IGF-I and IGFBP-3 concentrations may underlie some beneficial effects of vitamin D supplementation. However, it must be acknowledged that simultaneous changes in IGF-I and IGFBP-3 levels may lead to too little or too much free IGF-I being available for endocrine actions, which is anyway detrimental. This issue was recently brought up by a study which found a decline in the IGF-I/IGFBP-3 ratio with increasing 25(OH)D quartiles in severely obese patients and a decrease in the IGF-I/IGFBP-3 ratio following vitamin D treatment in overweight and nonseverely obese subjects.[48]

Conclusions

The interplay between vitamin D and IGF-I is complicated and occurs at both endocrine and paracrine/autocrine levels.

Vitamin D has been shown to increase circulating IGF-I and IGFBP-3, with the consistent finding of a positive correlation between 25(OH)D and IGF-I in population-based cohorts of healthy subjects. Preliminary results suggest that because of this effect, vitamin D status may have an impact on the titration of rhGH according to IGF-I concentrations in patients with GHD. Furthermore, the modulation of circulating IGF-I and IGFBP-3 might subtend some of the beneficial health effects ascribed to vitamin D. Therefore, we suggest that IGF-I and IGFBP-3 be taken into consideration in investigations of the health benefits of vitamin D.

By stimulating the synthesis of 1,25(OH)2D in the kidney, IGF-I promotes vitamin D endocrine activity. As a result, uncontrolled acromegaly is characterized by a tendency to hypercalcaemia, hypercalciuria and hyperphosphataemia. On the other hand, the induction of renal 1α-hydroxylase may partially account for the favourable modulation of bone metabolism by rhGH.

In spite of an abundance of experimental data, the meaning and importance of the paracrine/autocrine interactions between the vitamin D and IGF-I systems are still unclear. Breast cancer, in which vitamin D modulates the IGF-I/IGFBP ratio to decrease proliferation and increase apoptosis, represents an exception and may serve as a reference model for future studies.

Competing interests / financial disclosure

Nothing to declare.

Ancillary